An application of space filling curves to digital terrain models

An application of space filling curves to digital terrain models

A DTM (Digital Terrain Model) contains spatial elevation data of the terrain in a digital format and can find multiple applications in the planning and creation of physical models such as civil and mining engineering structures. A fundamental issue of dense DTMs is data compression. Compressing a DTM decreases the storage space and makes possible faster access to the data as well as faster transmission.

In this context, this work relates to the performance of scan patterns or space filling curves as applied to 16-bit DTM representations. These patterns are appropriate for describing terrain models in 1D while preserving the integrity of the elevation data they represent. In particular, this paper outlines the use of the so called space-filling curves for DTM compression.

The basic process of using 1D space filling curves for representing DTMs has previously been described but the application of the Peano-Hilbert (PH) fractal curve is the extension detailed here. The PH curve defines a continuous 1D scan that traverses through every height of a DTM exactly once, resulting in a representation that offers an improved neighbour height similarity.

The increased similarity allows for better results in terms of DTM coding/decoding times and compression ratios if compared, for instance, to existing digital image formats used to characterize DTMs. This is demonstrated by the results of a number of tests based on DTMs of various terrain characteristics and complexity (i.e. DTM entropy).

c. Ellipses Media.